Unsupervised Deep Video Denoising with Untrained Network

Deep learning has become a prominent tool for video denoising. However, most existing deep denoising methods require supervised training using noise-free videos. Collecting videos can be costly and challenging in many applications. Therefore, this paper aims to develop an unsupervised method that only uses single test noisy training. To achieve this, loss function is presented provides unbiased estimator of its counterpart defined on video. Additionally, temporal attention mechanism proposed exploit redundancy among frames. The experiments demonstrate the outperforms remains competitive against recent methods.